DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Status of Claims
Claims 1-12 are pending of which claims 1, 11, and 12 are in independent form.
Claim 11 is subject to claim interpretation.
Claims 1-12 are rejected under 35 U.S.C. 103.
Claim Interpretation
The following is a quotation of 35 U.S.C. 112(f):
(f) Element in Claim for a Combination. – An element in a claim for a combination may be expressed as a means or step for performing a specified function without the recital of structure, material, or acts in support thereof, and such claim shall be construed to cover the corresponding structure, material, or acts described in the specification and equivalents thereof.
The following is a quotation of pre-AIA 35 U.S.C. 112, sixth paragraph:
An element in a claim for a combination may be expressed as a means or step for performing a specified function without the recital of structure, material, or acts in support thereof, and such claim shall be construed to cover the corresponding structure, material, or acts described in the specification and equivalents thereof.
The claims in this application are given their broadest reasonable interpretation using the plain meaning of the claim language in light of the specification as it would be understood by one of ordinary skill in the art. The broadest reasonable interpretation of a claim element (also commonly referred to as a claim limitation) is limited by the description in the specification when 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, is invoked.
As explained in MPEP § 2181, subsection I, claim limitations that meet the following three-prong test will be interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph:
(A) the claim limitation uses the term “means” or “step” or a term used as a substitute for “means” that is a generic placeholder (also called a nonce term or a non-structural term having no specific structural meaning) for performing the claimed function;
(B) the term “means” or “step” or the generic placeholder is modified by functional language, typically, but not always linked by the transition word “for” (e.g., “means for”) or another linking word or phrase, such as “configured to” or “so that”; and
(C) the term “means” or “step” or the generic placeholder is not modified by sufficient structure, material, or acts for performing the claimed function.
Use of the word “means” (or “step”) in a claim with functional language creates a rebuttable presumption that the claim limitation is to be treated in accordance with 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph. The presumption that the claim limitation is interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, is rebutted when the claim limitation recites sufficient structure, material, or acts to entirely perform the recited function.
Absence of the word “means” (or “step”) in a claim creates a rebuttable presumption that the claim limitation is not to be treated in accordance with 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph. The presumption that the claim limitation is not interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, is rebutted when the claim limitation recites function without reciting sufficient structure, material or acts to entirely perform the recited function.
Claim limitations in this application that use the word “means” (or “step”) are being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, except as otherwise indicated in an Office action. Conversely, claim limitations in this application that do not use the word “means” (or “step”) are not being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, except as otherwise indicated in an Office action.
This application includes one or more claim limitations that do not use the word “means,” but are nonetheless being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, because the claim limitation(s) uses a generic placeholder that is coupled with functional language without reciting sufficient structure to perform the recited function and the generic placeholder is not preceded by a structural modifier. Such claim limitation(s) is/are:
“a sensor parameter candidate determination unit…”; “a sensor data generation unit” ; “a neural network model training unit”; “a calculation unit”; “ a selection unit ”; and “an output unit” in claim 11.
The corresponding structure is described:
“a sensor parameter candidate determination unit” ¶ [0051]-[0055], [0108]-[0109], [0121]-[0123], and [0141] of the specification, provides corresponding structure/algorithm. 35 USC 112(b) stratified.
“a sensor data generation unit” ¶ [0056]-[0058], [0094]-[0103], [0110]-[0111], and [0141] of the specification, provides corresponding structure/algorithm. 35 USC 112(b) stratified.
“a neural network model training unit” ¶ [0059]-[0061], [0112]-[0115] of the specification, provides corresponding structure/algorithm. 35 USC 112(b) stratified.
“a calculation unit” ¶ [0062]-[0063], [0116] of the specification, provides corresponding structure/algorithm. 35 USC 112(b) stratified.
“a selection unit” ¶ [0073], [0132] of the specification, provides corresponding structure/algorithm. 35 USC 112(b) stratified.
“an output unit” ¶ [0133]-[0136] of the specification, provides corresponding structure/algorithm. 35 USC 112(b) stratified.
Therefore, regarding Claim 11 is not subject to 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, because the limitation “unit…” (placeholder) recites corresponding structures/algorithms performing a list of purely functioning operations, reciting corresponding structure in the claims/specification.
Because this/these claim limitation(s) is/are being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, it/they is/are being interpreted to cover the corresponding structure described in the specification as performing the claimed function, and equivalents thereof.
If applicant does not intend to have this/these limitation(s) interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, applicant may: (1) amend the claim limitation(s) to avoid it/them being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph (e.g., by reciting sufficient structure to perform the claimed function); or (2) present a sufficient showing that the claim limitation(s) recite(s) sufficient structure to perform the claimed function so as to avoid it/them being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph.
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claim(s) 1, 2, 6, and 10-12 are rejected under 35 U.S.C. 103 as being unpatentable over SATO; Satoshi et al. (US 20190347520 A1) [Sato] in view of Das; Debkalpo (US 20220207223 A1) [Das] in view of Maher; Joshua Thomas (US 20230059313 A1) [Maher].
Regarding claims 1, 11, and 12, Sato discloses, a training method, by a computer, comprising: determining a plurality of sensor parameter candidates that are candidates for sensor parameters to be used for an operation of a sensor (Light-field camera 211 illustrated in FIG. 5 includes multi-pinhole mask 211a and image sensor 211b. Multi-pinhole mask 211a is placed a certain distance away from image sensor 211b. Multi-pinhole mask 211a includes a plurality of randomly or evenly arranged pinholes 211aa. The plurality of pinholes 211aa are also called multiple pinholes. Image sensor 211b obtains the image of a subject through each of the plurality of pinholes 211aa ¶ [0095], [0131], also see ¶ [0060]-[0069]);
and including sensor data to be obtained by the operation of the sensor (imager 11 as hardware includes a camera, and more specifically, a multi-pinhole camera, a coded aperture camera, a light-field camera, or a lensless camera. Such cameras enable imager 11 to simultaneously obtain a plurality of images of a subject by performing an operation for capturing an image one time, as will be described later. Note that imager 11 may obtain the plurality of images mentioned above by performing an operation for capturing an image multiple times by, for example, changing the imaging area that is the light-receiving area of an image sensor included in imager 11. Imager 11 outputs a second computational imaging image obtained to obtainer 101 in image identification device 10 ¶ [0060], [0063]-[0065], [0067], [0095]) and a plurality of pieces of correct answer identification information corresponding to each of the sensor data (Correct identification obtainer 123 obtains correct identification for machine learning using a first computational imaging image obtained by first image obtainer 121. The correct identification may be provided from outside of identification system 1A together with the first computational imaging image, or a user may input, for example, manually to provide the correct identification ¶ [0087]-[0089], [0100], [0103]-[0109], [0082]);
and training the neural network model by using an error between an identification result output from the neural network model and the correct answer identification information corresponding to the input some sensor data (Trainer 124 trains a classifier used by identifier 102 using a first computational imaging image obtained by first image obtainer 121 and correct identification obtained by correct identification obtainer 123 which corresponds to a captured image that is obtained by second image obtainer 122. Trainer 124 causes the classifier stored in second memory 224 to perform machine learning, and stores the latest classifier that has been trained into second memory 224. Identifier 102 obtains the latest classifier stored in second memory 224, stores the latest classifier into first memory 203, and uses the latest classifier for identification processing. The above machine learning is realized using backpropagation (BP) and the like in deep learning, for example. More specifically, trainer 124 inputs the first computational imaging image to the classifier, and obtains an identification result which the classifier outputs. Then, trainer 124 adjusts the classifier such that the identification result is to be the correct identification. Trainer 124 improves identification accuracy of the classifier by repeating such adjustment to a plurality of first computational imaging images, each of which is different, and a plurality of correct identification that correspond to each of the plurality of first computational imaging images (for example, thousands of pairs) ¶ [0089]; examiner specifies that the training classifier by comparing the identification results with the correct identification information and repeatedly adjusting the classier until the identification results become the correct identification);
and the sensor parameter candidate corresponding to the trained neural network model candidate with the highest identification performance (individualized ML corresponding to different camera (sensor) configurations (such as different pinhole positions and sizes). Therefore, selection of the highest-performing trained model inherently identifies the corresponding sensor parameter configuration used to generate that model ¶ [0131]);
outputting the selected pair of the sensor parameter candidate and the trained neural network model candidate (A great number of pairs (for example, thousands of pairs) of a captured image as illustrated in FIG. 6 and a first computational imaging image as illustrated in FIG. 7 are to be prepared. Trainer 124 obtains a classifier stored in second memory 224, obtains an output result obtained by inputting the first computational imaging image to the classifier, and adjusts the classifier such that the output result is to be correct identification obtained by inputting the captured image which corresponds to the first computational imaging image. Then, trainer 124 updates the classifier in second memory 224 by storing the adjusted classifier into second memory 224 ¶ [0109]-[0110]; also see ¶ [0112]-[0114]).
However, Sato does not explicitly facilitate calculating identification performance of the plurality of trained neural network model candidates; by using another sensor data of the sensor data included in the sensor data sets and the correct answer identification information corresponding to the other sensor data; selecting a pair of the trained neural network model candidate with the highest identification performance.
Das discloses, calculating identification performance of the plurality of trained neural network model candidates (The validation engine 184 can determine an accuracy of virtual model 190 based on the corresponding sets of features of the validation set. The validation engine 184 can discard a trained virtual model 190 that has an accuracy that does not meet a threshold accuracy. In some embodiments, the selection engine 185 can be capable of selecting a trained virtual model 190 that has an accuracy that meets a threshold accuracy. In some embodiments, the selection engine 185 can be capable of selecting the trained virtual model 190 that has the highest accuracy of the trained virtual models 190 ¶ [0036]) by using another sensor data of the sensor data included in the sensor data sets and the correct answer identification information corresponding to the other sensor data (Server machine 170 includes a training set generator 172 and a physics engine 174 that is capable of generating training data sets (e.g., a set of data inputs and a set of target outputs) to train, validate, and/or test a virtual model 190. Virtual model 190 can be a statistics based virtual model, a machine learning model, or any other algorithmic model capable of learning from data ¶ [0033]; The validation engine 184 can be capable of validating virtual model 190 using a corresponding set of features of a validation set from training set generator 172. The validation engine 184 can determine an accuracy of virtual model 190 based on the corresponding sets of features of the validation set. The validation engine 184 can discard a trained virtual model 190 that has an accuracy that does not meet a threshold accuracy. In some embodiments, the selection engine 185 can be capable of selecting a trained virtual model 190 that has an accuracy that meets a threshold accuracy. In some embodiments, the selection engine 185 can be capable of selecting the trained virtual model 190 that has the highest accuracy of the trained virtual models 190 ¶ [0036]-[0037]);
selecting a pair of the trained neural network model candidate with the highest identification performance (The validation engine 184 can be capable of validating virtual model 190 using a corresponding set of features of a validation set from training set generator 172. The validation engine 184 can determine an accuracy of virtual model 190 based on the corresponding sets of features of the validation set. The validation engine 184 can discard a trained virtual model 190 that has an accuracy that does not meet a threshold accuracy. In some embodiments, the selection engine 185 can be capable of selecting a trained virtual model 190 that has an accuracy that meets a threshold accuracy. In some embodiments, the selection engine 185 can be capable of selecting the trained virtual model 190 that has the highest accuracy of the trained virtual models 190 ¶ [0036]-[0037]; examiner specifies that this is selecting model with the highest accuracy);
It would have been obvious to one ordinary skilled in the art at the time of the present invention to combine the teachings of the cited references because Das' system would have allowed Sato to facilitate calculating identification performance of the plurality of trained neural network model candidates; by using another sensor data of the sensor data included in the sensor data sets and the correct answer identification information corresponding to the other sensor data; selecting a pair of the trained neural network model candidate with the highest identification performance. The motivation to combine is apparent in the Sato’s reference, because there is a need to improve capability of generating cost and time efficient metrology data.
However, neither Maher nor Das explicitly facilitates generating a plurality of sensor data sets corresponding to each of the plurality of sensor parameter candidates; generating a plurality of trained neural network model candidates corresponding to the plurality of sensor parameter candidates by inputting some of the sensor data included in each of the plurality of sensor data sets into a neural network model corresponding to the sensor data set.
Maher discloses, generating a plurality of sensor data sets corresponding to each of the plurality of sensor parameter candidates (generating multiple training, validation and testing datasets corresponding to different sensor subsets and feature sets, including sensor data associated with respective sensor configuration ¶ [0051]-[0052]; also see ¶ [0071], [0083]-[0086]);
generating a plurality of trained neural network model candidates corresponding to the plurality of sensor parameter candidates by inputting some of the sensor data included in each of the plurality of sensor data sets into a neural network model corresponding to the sensor data set (The training engine 182 may generate multiple trained machine learning models 190, where each trained machine learning model 190 corresponds to a distinct set of features of the training set (e.g., sensor data from a distinct set of sensors). For example, a first trained machine learning model may have been trained using all features (e.g., X1-X5), a second trained machine learning model may have been trained using a first subset of the features (e.g., X1, X2, X4), and a third trained machine learning model may have been trained using a second subset of the features (e.g., X1, X3, X4, and X5) that may partially overlap the first subset of features. Data set generator 172 may receive the output of a trained machine learning model (e.g., 190A), collect that data into training, validation, and testing data sets, and use the data sets to train a second machine learning model (e.g., 190B) ¶ [0052]; At block 312B, the system 300B performs model training (e.g., via training engine 182 of FIG. 1) using the training set 302B. The system 300B may train multiple models using multiple sets of features of the training set 302B (e.g., a first set of features of the training set 302B, a second set of features of the training set 302B, etc.). For example, system 300B may train a machine learning model to generate a first trained machine learning model using the first set of features in the training set (e.g., sensor data from sensors 1-10 for products 1-60) and to generate a second trained machine learning model using the second set of features in the training set (e.g., sensor data from sensors 11-20 for products 1-60) ¶ [0085]-[0086], [0088]; also see ¶ [0049]-[0051]).
It would have been obvious to one ordinary skilled in the art at the time of the present invention to combine the teachings of the cited references because Maher's system would have allowed Sato and Das to facilitate generating a plurality of sensor data sets corresponding to each of the plurality of sensor parameter candidates; generating a plurality of trained neural network model candidates corresponding to the plurality of sensor parameter candidates by inputting some of the sensor data included in each of the plurality of sensor data sets into a neural network model corresponding to the sensor data set. The motivation to combine is apparent in the Sato and Das’ reference, because there is a need to improve dimensionality reduction.
Regarding claim 2, the combination of Sato, Maher and Das discloses, wherein the sensor is a multi-pinhole camera including a multi-pinhole mask in which a plurality of pinholes is formed and an image sensor, and the sensor parameters are at least one of a distance between the multi-pinhole mask and the image sensor, a number of the plurality of pinholes, a size of each of the plurality of pinholes, and a position of each of the plurality of pinholes (Sato: Light-field camera 211 illustrated in FIG. 5 includes multi-pinhole mask 211a and image sensor 211b. Multi-pinhole mask 211a is placed a certain distance away from image sensor 211b. Multi-pinhole mask 211a includes a plurality of randomly or evenly arranged pinholes 211aa. The plurality of pinholes 211aa are also called multiple pinholes. Image sensor 211b obtains the image of a subject through each of the plurality of pinholes 211aa. The image obtained through a pinhole is called a pinhole image. Since the subject included in each of pinhole images differs depending on the position and the size of each pinhole 211aa, image sensor 211b obtains a superimposed image in which the subject is superimposed multiple times. The position of pinhole 211aa affects the position of a subject projected onto image sensor 211b, and the size of pinhole 211aa affects a blur of a pinhole image. By using multi-pinhole mask 211a, it is possible to obtain a superimposed image in which a subject included in each of a plurality of pinhole images which is captured in different positions and the degree of blur is different is superimposed multiple times. When a subject is away from pinhole 211aa, the subject included in each of a plurality of pinhole images is projected at almost the same position. On the contrary, when the subject is close to pinhole 211aa, the subjects each of which is included in each of a plurality of pinhole images are projected separately. Since the amount of displacement of the subjects each included in each of the plurality of pinhole images which is superimposed multiple times corresponds to a distance between the subject and multi-pinhole mask 211a, a superimposed image includes depth information of the subject according to the amount of displacement ¶ [0095]; also see ¶ [0131]).
Regarding claim 6, the combination of Sato, Maher and Das discloses, wherein generating the plurality of sensor data sets includes generating the sensor data by generating a plurality of images captured from a plurality of virtual viewpoint positions based on the sensor parameters by computer graphics, and superimposing the plurality of generated images (Sato: As described above, computational imaging images (a first computational imaging image and a second computational imaging image) are images each of which includes parallax information indicating that an object and the surrounding environment of the object are superimposed multiple times ¶ [0043], [0060], [0067]; The image obtained through a pinhole is called a pinhole image. Since the subject included in each of pinhole images differs depending on the position and the size of each pinhole 211aa, image sensor 211b obtains a superimposed image in which the subject is superimposed multiple times. The position of pinhole 211aa affects the position of a subject projected onto image sensor 211b, and the size of pinhole 211aa affects a blur of a pinhole image. By using multi-pinhole mask 211a, it is possible to obtain a superimposed image in which a subject included in each of a plurality of pinhole images which is captured in different positions and the degree of blur is different is superimposed multiple times. When a subject is away from pinhole 211aa, the subject included in each of a plurality of pinhole images is projected at almost the same position. On the contrary, when the subject is close to pinhole 211aa, the subjects each of which is included in each of a plurality of pinhole images are projected separately. Since the amount of displacement of the subjects each included in each of the plurality of pinhole images which is superimposed multiple times corresponds to a distance between the subject and multi-pinhole mask 211a, a superimposed image includes depth information of the subject according to the amount of displacement ¶ [0095], [0098]. The image sensor simultaneously obtains a plurality of images through the pluralities of pinholes or microlenses by performing operation for capturing an image one time among the whole operations. The plurality of images are captured from different viewpoints. From the positional relationship between the plurality of images and the viewpoints, it is possible to obtain a range of a subject in a depth direction ¶ [0064]).
Regarding claim 10, the combination of Sato, Maher and Das discloses, wherein the sensor is a coded aperture camera including a coded mask in which a plurality of pinholes is formed and an image sensor, and the sensor parameters are at least one of a distance between the coded mask and the image sensor, a number of the plurality of pinholes, a size of each of the plurality of pinholes, and a position of each of the plurality of pinholes (Sato: In addition, although the above has described imager 11 as a light-field camera which uses multiple pinholes or a microlens, imager 11 is not limited to the above. For example, imager 11 may be a configuration which captures a coded aperture image. This configuration is also a type of multi-pinhole camera ¶ [0123]. FIG. 12 is a schematic diagram of an example of a coded aperture mask that uses a random mask as a coded diaphragm ¶ [0025]. computational imaging image include coded images, such as a light-field image using multiple pinholes or a microlens, a compressed sensing image captured by performing weighting addition on pixel information in time and space, and a coded aperture image captured using a coded aperture and a coded aperture mask ¶ [0063]. Also see ¶ [0065], [0095], [0124], [0125], and [0131]).
Claim(s) 3-5 are rejected under 35 U.S.C. 103 as being unpatentable over Sato in view of Das in view of Maher in view of Aswin; Buddy (US 20190122378 A1) [Aswin].
Regarding claim 3, the combination of Sato, Maher and Das discloses, wherein generating the plurality of sensor data sets [includes generating the sensor data by performing a process of convolving a point spread function corresponding to] the sensor parameters with the sensor data obtained by one of the pinholes (Sato: Light-field camera 211 illustrated in FIG. 5 includes multi-pinhole mask 211a and image sensor 211b. Multi-pinhole mask 211a is placed a certain distance away from image sensor 211b. Multi-pinhole mask 211a includes a plurality of randomly or evenly arranged pinholes 211aa ¶ [0095], also see ¶ [0017], [0019], [0064]).
However neither one of Sato, Maher, or Das explicitly facilitates includes generating the sensor data by performing a process of convolving a point spread function corresponding to.
Aswin discloses, includes generating the sensor data by performing a process of convolving a point spread function corresponding to (Generally, in Step 509, the sharper image can be convolved with a range of PSF filters that may match an amount of blur in a defocused image. An exemplary PSF filter set can be a list of 2D Gaussian transfer function(s) with different variance and total 2D integration value of one ¶ [0075], [0079]. at this step determine camera calibration and correct optical aberrations in 2D images output from camera (e.g., pixel values, coordinates, etc)) settings for imager shape and orientation, characterize point spread function, magnification and other non-ideal image characteristics for each camera settings (e.g. e.g., focus, aperture opening/size, exposure, zoom, etc)). Step 201 Output: Unprocessed 2D images data structure (DS) 1001 ¶ [0037]-[0038]. Also see ¶ [0050]).
It would have been obvious to one ordinary skilled in the art at the time of the present invention to combine the teachings of the cited references because Aswin's system would have allowed Sato, Das and Maher to facilitate includes generating the sensor data by performing a process of convolving a point spread function corresponding to. The motivation to combine is apparent in the Sato, Das and Maher’s reference, because there a need to improve machine vision systems used to create models based on two dimensional images from multiple perspectives, multiple camera settings.
Regarding claim 4, the combination of Sato, Maher, Das, and Aswin discloses, wherein the sensor is a multi-pinhole camera including a multi-pinhole mask in which a plurality of pinholes is formed and an image sensor (Sato: Light-field camera 211 illustrated in FIG. 5 includes multi-pinhole mask 211a and image sensor 211b. Multi-pinhole mask 211a is placed a certain distance away from image sensor 211b. Multi-pinhole mask 211a includes a plurality of randomly or evenly arranged pinholes 211aa ¶ [0095], also see ¶ [0131]), and the sensor parameters are at least one of a scaling parameter, a rotation parameter, and a skew parameter for performing affine transformation on the plurality of entire pinholes (Aswin: Translation and rotation of a camera … image shifting and scaling ¶ [0006]; Example can include an input image that is distorted (e.g., skewed, stretched, etc) and an output can include undistorted output ¶ [0038]; a scaling constant ¶ [0056]; the scaling variable ¶ [0057]; also see ¶ [0069], [0079]).
Regarding claim 5, the combination of Sato, Maher, Das, and Aswin discloses, wherein generating the plurality of sensor data sets includes (Sato: Light-field camera 211 illustrated in FIG. 5 includes multi-pinhole mask 211a and image sensor 211b. Multi-pinhole mask 211a is placed a certain distance away from image sensor 211b. Multi-pinhole mask 211a includes a plurality of randomly or evenly arranged pinholes 211aa ¶ [0095]), generating the sensor data by performing a process of convolving a point spread function (Aswin: Generally, in Step 509, the sharper image can be convolved with a range of PSF filters that may match an amount of blur in a defocused image. An exemplary PSF filter set can be a list of 2D Gaussian transfer function(s) with different variance and total 2D integration value of one ¶ [0074]-[0075], [0079], [0050]) on which the affine transformation is performed according to the sensor parameters with the sensor data obtained by one of the pinholes (Aswin: Note that Step 201 and subsequent steps can be implemented using other camera models (e.g. Plenoptic camera, fish eye camera, etc) beside a pin-hole camera, can be used providing that proper calibration and projective transformation are performed ¶ [0069], [0079]).
Claim(s) 7-9 are rejected under 35 U.S.C. 103 as being unpatentable over Sato in view of Das in view of Maher in view of Berkenkamp; Felix et al. (US 20220097227 A1) [Berkenkamp].
Regarding claim 7, the combination of Sato, Maher, and Das teaches all the limitations of claim 1.
However neither one of Sato, Maher, or Das explicitly facilitates wherein determining the plurality of sensor parameter candidates includes determining the plurality of sensor parameter candidates by black box optimization based on the plurality of sensor parameter candidates previously determined and the identification performance of the plurality of trained neural network model candidates corresponding to each of the plurality of sensor parameter candidates.
Berkenkamp discloses, wherein determining the plurality of sensor parameter candidates includes determining the plurality of sensor parameter candidates by black box optimization based on the plurality of sensor parameter candidates previously determined and the identification performance of the plurality of trained neural network model candidates corresponding to each of the plurality of sensor parameter candidates (Bayesian optimization (BO) is suitable in such cases for ascertaining control parameter values since it allows unknown (“black box”) functions ¶ [0003]; A neural network (NN) is used with this method to encode a feature space which is jointly used by all previously examined black box functions (which are related to the present task). One Bayesian linear regression (BLR) layer per task is used to learn a representation of the task, including the expected value for a given set of parameters and the uncertainty with respect to this value. By learning the shared feature space, MT-ABLR is able to transfer available knowledge from similar black box functions, and to carry out the optimization for the present task more efficiently ¶ [0005]; a task for processing sensor data (e.g., image classification) stemming from similar sensors (e.g., from another camera type) ¶ [0008]; The above-described approach is also suitable for rapidly finding hyperparameters for a machine learning model, for example for a neural network for processing sensor data ¶ [0023]; Since the relationship between control configuration and result is complex and difficult to predict, i.e., is given by an unknown “black box function,” control unit 106 determines the control configuration with the aid of Bayesian optimization ¶ [0043], [0054]).
It would have been obvious to one ordinary skilled in the art at the time of the present invention to combine the teachings of the cited references because Berkenkamp's system would have allowed Sato, Das and Maher to facilitate wherein determining the plurality of sensor parameter candidates includes determining the plurality of sensor parameter candidates by black box optimization based on the plurality of sensor parameter candidates previously determined and the identification performance of the plurality of trained neural network model candidates corresponding to each of the plurality of sensor parameter candidates. The motivation to combine is apparent in the Sato, Das and Maher’s reference, because there a need to improve carrying out the optimization for the task more efficiently.
Regarding claim 8, the combination of Sato, Maher, Das, Berkenkamp discloses, wherein determining the plurality of sensor parameter candidates includes determining the plurality of sensor parameter candidates by black box optimization based on the plurality of sensor parameter candidates previously determined, the identification performance of the plurality of trained neural network model candidates corresponding to each of the plurality of sensor parameter candidates, and an index indicating confidentiality of the sensor data (Berkenkamp: Bayesian optimization (BO) is suitable in such cases for ascertaining control parameter values since it allows unknown (“black box”) functions ¶ [0003]; A neural network (NN) is used with this method to encode a feature space which is jointly used by all previously examined black box functions (which are related to the present task). One Bayesian linear regression (BLR) layer per task is used to learn a representation of the task, including the expected value for a given set of parameters and the uncertainty with respect to this value. By learning the shared feature space, MT-ABLR is able to transfer available knowledge from similar black box functions, and to carry out the optimization for the present task more efficiently ¶ [0005]; a task for processing sensor data (e.g., image classification) stemming from similar sensors (e.g., from another camera type) ¶ [0008]; The above-described approach is also suitable for rapidly finding hyperparameters for a machine learning model, for example for a neural network for processing sensor data ¶ [0023]; Since the relationship between control configuration and result is complex and difficult to predict, i.e., is given by an unknown “black box function,” control unit 106 determines the control configuration with the aid of Bayesian optimization ¶ [0043], [0054]) and an index indicating confidentiality of the sensor data (Maher: the data set generator 272 may discretize (e.g., segment) one or more of the data input 210 or the target output 220 (e.g., to use in classification algorithms for regression problems). Discretization (e.g., segmentation via a sliding window) of the data input 210 or target output 220 may transform continuous values of variables into discrete values. In some embodiments, the discrete values for the data input 210 indicate discrete historical sensor data 244 to obtain a target output 220 (e.g., discrete compressed data 230B) ¶ [0072]).
Regarding claim 9, the combination of Sato, Maher, Das, Berkenkamp discloses, wherein the black box optimization is Bayesian estimation (Berkenkamp: Bayesian optimization (BO) is suitable in such cases for ascertaining control parameter values since it allows unknown (“black box”) functions ¶ [0003]; A neural network (NN) is used with this method to encode a feature space which is jointly used by all previously examined black box functions (which are related to the present task). One Bayesian linear regression (BLR) layer per task is used to learn a representation of the task, including the expected value for a given set of parameters and the uncertainty with respect to this value. By learning the shared feature space, MT-ABLR is able to transfer available knowledge from similar black box functions, and to carry out the optimization for the present task more efficiently ¶ [0005]; a task for processing sensor data (e.g., image classification) stemming from similar sensors (e.g., from another camera type) ¶ [0008]; Since the relationship between control configuration and result is complex and difficult to predict, i.e., is given by an unknown “black box function,” control unit 106 determines the control configuration with the aid of Bayesian optimization ¶ [0043], [0054]).
Conclusion
The examiner requests, in response to this Office action, support be shown for language added to any original claims on amendment and any new claims. That is, indicate support for newly added claim language by specifically pointing to page(s) and line no(s) in the specification and/or drawing figure(s). This will assist the examiner in prosecuting the application.
When responding to this office action, Applicant is advised to clearly point out the patentable novelty which he or she thinks the claims present, in view of the state of the art disclosed by the references cited or the objections made. He or she must also show how the amendments avoid such references or objections See 37 CFR 1.111(c).
Any inquiry concerning this communication or earlier communications from the examiner should be directed to MOHAMMAD S ROSTAMI whose telephone number is (571)270-1980. The examiner can normally be reached Mon-Fri From 9 a.m. to 5 p.m..
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Boris Gorney can be reached at (571)270-5626. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
8/16/2026
/MOHAMMAD S ROSTAMI/ Primary Examiner, Art Unit 2154
Prosecution Insights